2

The server seems not be able to support almost two or three hundred people visiting at the same time. We upload some new codes in Nov. 29the and people saw 404 Not Found frequently ever since that day. First we thought that just because there are too many people trying to visit at the same time because Nov. 30 is a deadline. But now about a week later, the server are still not stable and the "404 Not Found" are still occured for the most of the time. Even in early morning, when I'm sure that few people would get up that early and visit the site, I still could not visit the server. I think there must be something wrong in our configuration or in our code then.

We used prefork for our mpm module ( we have tried worker.c but things get even worse so we change it back to prefork). I can't remeber the detail configuration right now but I'm sure that the MaxClient and the Serverlimit are both set up to 600 ( we actually have only less than 500 year 4 students.). And it seems that the Keep Alive has been set to 15 seconds, the MaxRequestsPerChild has been set to 100000.

Besides, I remeber that when we first upload some other information for people to view and choose, the traffic is still busy but nothing bad happened then. So maybe it's because something happened in the new code we upload on Nov. 29. Does anybody know what operation would caause that problem when the code is executed or other reasons1?

1 Answers1

0

If users are getting 404 Not Found, it's not about Apache performance or sockets or similar.

If Apache can't open connection (reaching for example MaxClient limit or can't open socket) then users can't connect at all. If fastcgi (or whatever module is used to connect to Django) fails, error is 500 Internal Server error.

Are you absolutely sure your code uploads didn't include for example misplaced

raise Http404
Olli
  • 768
  • 6
  • 16